Macroeconomic Policy Outlook Q3 2020

This is our regular Macro Policy Outlook, providing a policy-focused take on the economy. In this edition, we focus on the labour market, and the prospects for unemployment.

While the OBR forecast that the unemployment rate will rise to match its 1980s peak of 11.9 per cent, the Bank of England is projecting a rise to ‘just’ 7.5 per cent, a smaller rise than seen during the financial crisis. So which of the two is more consistent with the data so far? The OBR’s forecast reflects its more pessimistic view about the hit to the economy, and an assumption that there will be minimal labour hoarding by firms.  On both issues, data released since that forecast provide some comfort: the economy has so far bounced back more quickly than expected, and we are not yet seeing the sharp rise in productivity that would be indicative of firms shedding labour (particularly in lower-paid, lower-productivity sectors). But, despite the bounce back through to July, the economy looks set to be operating well below pre-crisis levels when the Coronavirus Job Retention Scheme (JRS) ends next month. And even based on the Bank of England forecast, more than a million jobs will be at risk by that point. On top of this, the speed of the recovery over the summer implied by mobility data, as well as the increased use of localised restrictions, suggest that the recovery could be slower than envisaged by the Bank. Given this uncertainty, it is crucial that policy makers respond to changes in the outlook, acting to minimise the risk of a large, painful rise in unemployment.  

The Bank of England and OBR have very different outlooks for unemployment and who is right will matter a lot

Increases in official unemployment so far have been small, but that is set to change. The unemployment rate ticked up to 4.1 per cent in the three months to July – just 0.3 percentage points off its more-than-45-year low reached in January. But little comfort can be taken from this as it reflects the massive support to the labour market coming from the JRS that has kept workers attached to their jobs and counted as employed even when social distancing restrictions have made their work impossible or unprofitable. With that support set to be withdrawn by the end of October, a key question concerns the size of the future rise in unemployment. This is important not just because of the impacts of higher unemployment on family finances and well-being, but also because it is an important determinant of the nature of the long-term scarring effects on the economy. In this edition of the Macro Policy Outlook we consider what the data so far can tell us about the future rise in unemployment.

The outlook for unemployment is definitely very uncertain, and the Bank of England and OBR are forecasting very different outcomes. As shown in Figure 1, the OBR’s central scenario from its July Fiscal Sustainability Report features a rise in the unemployment rate to 11.9 per cent in the fourth quarter of this year when the JRS ends and workers are laid off, an amount equal to the previous high in the 1980s. By contrast, in its August Monetary Policy Report (MPR) the Bank of England forecasts the unemployment rate will rise to ‘just’ 7.5 per cent, below the financial crisis peak of 8.5 per cent. So why is this difference so big?

Figure 1: The Bank of England and OBR have very different forecasts for the rise in unemployment

The outlook for unemployment will depend on the extent to which firms hang on to labour and the size of the economic hit

There are two important determinants of an unemployment forecast. The first is the amount of labour hoarding we see. Because it is costly to hire and fire employees, firms have an incentive to hold onto staff during bad times. Different assumptions about the amount of labour hoarding that takes place show up as differences in productivity (GDP per hour worked): more hoarding means a weaker outlook for productivity.

The second is the extent of the hit to GDP. The larger the impact of the coronavirus pandemic on output in affected sectors, the larger the falls in how much labour they employ. For both of these, we can see clear differences in the views taken by the Bank and the OBR. We consider these in turn.

The OBR thinks we are unlikely to see much labour hoarding…

As shown in Figure 2, the Bank of England has a weaker forecast for productivity, expecting it to evolve in a similar way to the financial crisis. The OBR, by contrast, envisages a sharp rise in productivity in the immediate aftermath of the crisis of a similar order of magnitude to that seen after the 1970s recession.  Underpinning these is an assumption by the Bank that assumes that the JRS works as intended and allows ‘most’ workers to return to their jobs as the economy recovers; firms are effectively hoarding labour with employment falling less than output. In the OBR’s central scenario, 15 per cent of workers flow from furlough into unemployment when the JRS closes. Because those sectors worst affected by the crisis tend to be low-paying, low-productivity ones, this implies a jump in overall productivity.

Figure 2: The Bank of England and OBR productivity forecasts are quite different

…but the good news is that – so far at least – the productivity data aren’t consistent with the OBR forecast

So is there any evidence so far a 1970s-like jump in productivity, consistent with a large rise in unemployment? Here, the aggregate data provide some comfort (see Figure 3): hours of work have fallen sharply by 17.5 per cent in the three months to July, but the fall in GDP is almost the same. That said, after falling initially, productivity growth ticked up sharply in July, so we should clearly be careful about reading too much into this at this stage. Indeed, the data available at this point are even more uncertain than usual given problems in measuring the economy during the pandemic, for example higher-than-usual non-response rates to ONS surveys (see Figure 11 here).

Figure 3: Productivity data so far seem consistent with a slowing in productivity

The OBR is also much more pessimistic about about the fall in GDP

What about the other key factor driving the rise in unemployment: the hit to GDP? Here, while the OBR is expecting GDP in 2020 to fall by 12.4 per cent, the largest fall in in over 300 years, the Bank is forecasting a fall of ‘just’ 9.5 per cent (still the weakest in nearly a century). The Bank’s forecast – which was made later with the benefit of additional data on the extent of the initial bounce back – is close to the consensus among all forecasters (which is for a fall of 10 per cent, according to the September HMT panel), while the OBR forecast for GDP now looks somewhat pessimistic – see Figure 4. This suggests that, when the OBR come to update its forecast in November, its forecast hit to GDP will be smaller.

Figure 4: GDP is set to be well below normal for the rest of the year

The strong bounce back so far is good news, but what matters for unemployment is the extent of the recovery when the JRS closes

One reason why unemployment had not risen by much, even in July, is that the JRS helped ensure that wage bills have fallen by proportionately more than output, on average, across the main sectors of the economy. Indeed, as shown in Figure 5, in the available data to June, the proportion of workers covered by the JRS has – for almost all sectors – been larger than falls in output. This clearly reduces firms’ costs, reducing the pressure on firms to lay workers off. But, as the JRS stops at the end of October, the proportion of the wage bill covered will fall to zero. If at that point output remains depressed, there is a risk of a very substantial rise in unemployment.

Figure 5: For most sectors, the JRS has meant wage bills have fallen proportionately more than output

A key question then, is the extent to which the economy will have recovered by the time the furlough scheme ends. The Bank of England forecasts that GDP will be 5 per cent below its peak by that point (see Figure 4). To get a sense of how sectoral output developments might interact with the ending of the JRS, we undertake a simple mapping of the Bank of England’s forecast into what that implies for individual sectors’ falls in output. If we assume that losses in the number of jobs eligible for furlough are proportionate to those output losses, then around 1.25 million jobs would be lost by this point, with those jobs concentrated in three sectors: hospitality, retail and transport.

…and, looking ahead, there are signs that the recovery could slow from here, particularly if local lockdowns become more widespread

More worryingly, real-time data suggest that the improvement in economic activity has stalled. As shown in Figure 6, although Google mobility data shows a rapid return of activity in the retail and recreational sectors, the increase in workplace mobility appears to have slowed. Looking ahead, the experience of other countries suggests that retail and recreational mobility will settle somewhat below pre-crisis levels (see left-hand panel in Figure 6).  Although it is not straightforward to link such data to aggregate GDP, the simple mapping used in our previous work implies that we would need to see further substantial improvements in workplace mobility data to be consistent with the Bank of England forecasts (see right-hand panel). And with the virus caseload rising and localised restrictions on social contact and some aspects of economic activity increasing in scope, it seems bold to assume that mobility will carry on rising at its previous rate. All this suggests there may be limits to the extent that output can recover, particularly if there are further lockdowns.

Figure 6: Improvements in mobility have stalled as social distancing restrictions have been reintroduced

Policy makers will need to respond to changes in the outlook

Although there is a high degree of uncertainty about the outlook for unemployment, this not an argument for policy inaction. While the rise in unemployment may not be as severe as suggested by the OBR forecast, we are still on track for a rise in unemployment of over a million people even if the economy evolves in line with the Bank of England’s forecast. And with local lockdowns becoming more prevalent, there is a risk that the economy could still be operating well below normal levels when the JRS ends. The role of policy in this situation is to respond decisively to these macroeconomic drivers of the labour market and provide additional support if the outlook deteriorates.